3 ms·

The frequency matters because it would give a far better insight into expected boost behavior & power consumption. For example when you look up the 15w i5-1035

The frequency matters because it would give a far better insight into expected boost behavior & power consumption. For example when you look up the 15w i5-1035G1 and see 1ghz base, 3.6ghz boost, you can figure out that no, those geekbench results are not representative of the chip running at 15w. It's pulling way, way more because it's in the turbo window, and the gap between base & turbo is huuuuuge. So right now when Apple claims the M1 is 10w, we really have no context for that. Is it actually capped at 10w? Or is it 10w sustained like Intel's TDP measurements? How much faster is the M1 over the A14? How different is the M1's performance in the fanless Air vs. the fan MBP 13"? Frequency would give insights into most all of that. It's not useless.

giantrobot 6y ago
Here's the GeekBench scores for various iPhones [0] so you can answer for yourself. Here's the MacBook Air for comparison [1]. [0] https://browser.geekbench.com/ios_devices/iphone-12 https://browser.geekbench.com/ios_devices/iphone-12 [1] https://browser.geekbench.com/v5/cpu/4651056 https://browser.geekbench.com/v5/cpu/4651056
adrian_b 6y ago
The GeekBench tests have revealed that the clock frequency of the M1 cores is 3.2 GHz, both in MB Air and in MB Pro. Therefore M1 succeeds to have a slightly higher single-thread performance (between 3% and 8%) than Intel Tiger Lake and AMD Zen 3 at only 2/3 of their clock frequency. The speed advantage of M1 in single-thread is not high enough to be noticeable in practice, but what counts is that reaching the same performance at high IPC and low frequency instead of low IPC and high frequency results in having the same performance as the competitors at a much lower power consumption. This is the real achievement of Apple and not the ridiculous claims about M1 being 3 times faster than obsolete and slow older products.
- kllrnohj 6y ago
  Geekbench doesn't monitor the frequency over time. We don't know what the M1 actually ran at during the test, nor what it can boost to. Or if it has a boost behavior at all even.

giantrobot 6y ago

Here's the GeekBench scores for various iPhones [0] so you can answer for yourself. Here's the MacBook Air for comparison [1]. [0] https://browser.geekbench.com/ios_devices/iphone-12 https://browser.geekbench.com/ios_devices/iphone-12 [1] https://browser.geekbench.com/v5/cpu/4651056 https://browser.geekbench.com/v5/cpu/4651056

adrian_b 6y ago

The GeekBench tests have revealed that the clock frequency of the M1 cores is 3.2 GHz, both in MB Air and in MB Pro. Therefore M1 succeeds to have a slightly higher single-thread performance (between 3% and 8%) than Intel Tiger Lake and AMD Zen 3 at only 2/3 of their clock frequency. The speed advantage of M1 in single-thread is not high enough to be noticeable in practice, but what counts is that reaching the same performance at high IPC and low frequency instead of low IPC and high frequency results in having the same performance as the competitors at a much lower power consumption. This is the real achievement of Apple and not the ridiculous claims about M1 being 3 times faster than obsolete and slow older products.

kllrnohj 6y ago
Geekbench doesn't monitor the frequency over time. We don't know what the M1 actually ran at during the test, nor what it can boost to. Or if it has a boost behavior at all even.